A Mixed-Methods Study on the Effects of Misconceptions and Text Type on Metacomprehension

• In Event: Advances in Understanding Metacomprehension Accuracy of Expository Text

Abstract

Theory and Objectives: When learning from text, accurate monitoring of comprehension, which is known as metacomprehension accuracy, is considered the driving force for effective regulation of study and successful learning (Thiede et al., 2003). However, despite its importance, learners are often overconfident when monitoring their text comprehension (e.g., Maki et al., 2005). One factor that leads to overconfidence are misconceptions. Specifically, authors (2017) found that statistical misconceptions impaired not only learners’ comprehension of a statistics text but also their metacomprehension as indicated by overconfident predictions of comprehension. Consequently, it seems crucial to explore instructional methods that support learners in self-regulated learning from text when they hold misconceptions. A promising method to overcome the detrimental effects of misconceptions on comprehension are refutation texts. Refutation texts explicitly addresses typical misconceptions about a topic and directly refute them with the scientifically correct explanations (Tippett, 2010). Studies that investigated the cognitive processes taking place when reading refutation texts found that these texts lead to the coactivation of misconceptions and correct conceptions, thereby inducing conceptual-change processes like, for example, the experience of cognitive conflict and the evaluation of one’s current understanding (van den Broek & Kendeou, 2008). Yet, to what extent refutation texts are beneficial for metacomprehension besides of comprehension and to what extent conceptual-change processes during reading are linked to comprehension and metacomprehension after reading is not yet clear.

Method: To address these open questions, in the domain of statistics, we conducted an experiment with N = 80 university students and examined the reading processes of n = 26 of these students by using a think-aloud methodology. In the experiment, the participants first completed a misconceptions test about covariance. Then, they read a statistics text about covariance in either a standard or refutation text version. Next, the participants predicted the number of comprehension questions they would presumably answer correctly. Finally, they answered the comprehension questions. Metacomprehension accuracy was operationalized as the signed difference between judged and actual performance on the questions (absolute accuracy).


Results: Moderated mediation models revealed that the refutation text compensated for detrimental effects of misconceptions on text comprehension (see Table 1) and metacomprehension accuracy (see Table 2) by eliciting conceptual-change processes. More precisely, when reading the refutation text, students with a higher number of misconceptions showed a higher incident of conceptual-change processes. This, in turn, resulted in enhanced comprehension and less overconfidence. In contrast, when reading the standard text, this effect was not apparent. Instead, students’ misconceptions led to worse comprehension and greater overconfidence.

Significance: Apparently, the conceptual-change processes induced by reading the refutation text improved comprehension by replacing students’ flawed prior knowledge with the correct information provided in the text. At the same time, due to the induced conceptual-change processes, the students more closely evaluated their understanding (van den Broek & Kendeou, 2008) so as to provide more accurate predictions. In conclusion, in statistics as well as in other misconception-rich domains, refutation texts should increasingly be applied in instruction so as to foster self-regulated learning from text.

Authors

• Anja Prinz, University of Freiburg

• Stefanie Golke, University of Freiburg

• Joerg Wittwer, University of Freiburg